For the best experience we recommend upgrading to the latest version of these supported browsers:
I wish to continue viewing on my unsupported browser
For the best experience we recommend upgrading to the latest version of these supported browsers:
I wish to continue viewing on my unsupported browser
Tiffany alleged that eBay failed to take sufficient steps to prevent the sale of counterfeit Tiffany products on its website.
NERA, on behalf of eBay, was asked to evaluate the work performed by the Plaintiffs' expert who designed and selected a sample of Tiffany items listed on eBay. This sample was intended to provide an unbiased estimate of the extent to which listings of Tiffany goods for sale on eBay included counterfeit Tiffany & Co. merchandise. The NERA analysis exposed numerous flaws in both the sample design and its implementation. The sample was purportedly an equal probability sample, but NERA experts demonstrated that the sample was not, in fact, a probability sample in which every population element has a known, non-zero chance of selection. In order to permit the statistician to generalize from the sample to the larger population from which it was selected, a sample must be selected by probability methods. The problem that Plaintiffs' expert failed to consider is that items on eBay can be listed for different periods of time. As a result, the expert's selection method assigned higher probabilities to items that were listed for longer periods. Thus, instead of having an equal probability sample, Plaintiffs' expert had drawn a sample in which the probabilities of selection were unknown. The NERA experts not only explained the flaws in Plaintiffs' expert’s method, but also described the way in which a properly designed probability sample could have been created.
The court ruled against the Plaintiffs. As part of its decision, the court stated that the Plaintiffs’ estimate of the extent to which counterfeit Tiffany items are listed on eBay was not based on a probability sample, making it impossible to create an unbiased estimate or to calculate a confidence interval.